Psychoeducation As a Mediator of Treatment Approach on Parent Engagement in Child Psychotherapy for Disruptive Behavior

Journal of Clinical Child & Adolescent Psychology

ISSN: 1537-4416 (Print) 1537-4424 (Online) Journal homepage: http://www.tandfonline.com/loi/hcap20

Psychoeducation as a Mediator of Treatment Approach on Parent Engagement in Child Psychotherapy for Disruptive Behavior

Jonathan I. Martinez, Anna S. Lau, Bruce F. Chorpita, John R. Weisz & Research Network on Youth Mental Health

To cite this article: Jonathan I. Martinez, Anna S. Lau, Bruce F. Chorpita, John R. Weisz & Research Network on Youth Mental Health (2015): Psychoeducation as a Mediator of Treatment Approach on Parent Engagement in Child Psychotherapy for Disruptive Behavior, Journal of Clinical Child & Adolescent Psychology, DOI: 10.1080/15374416.2015.1038826

To link to this article: http://dx.doi.org/10.1080/15374416.2015.1038826

Published online: 04 Jun 2015.

Submit your article to this journal

Article views: 207

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hcap20

Download by: [Smithsonian Astrophysics Observatory] Date: 24 August 2016, At: 13:47 Journal of Clinical Child & Adolescent Psychology, 0(0), 1–15, 2015 Copyright # Taylor & Francis Group, LLC ISSN: 1537-4416 print=1537-4424 online DOI: 10.1080/15374416.2015.1038826

Psychoeducation as a Mediator of Treatment Approach on Parent Engagement in Child Psychotherapy for Disruptive Behavior

Jonathan I. Martinez Department of Psychology, University of California, Los Angeles Child & Adolescent Services Research Center

Anna S. Lau and Bruce F. Chorpita Department of Psychology, University of California, Los Angeles

John R. Weisz Department of Psychology, Harvard University

Research Network on Youth Mental Health Chicago, Illinois

Parent engagement in treatment for child disruptive behavior has been associated with improved child outcomes in care. However, many families who enter care do not receive an adequate dose of treatment, and parents are often not involved. We examined therapists’ use of psychoeducation, a therapeutic practice used to present factual information about target problems and treatments, and its association with parent engagement in child psychotherapy. Participants were drawn from the Child System and Treatment Enhancement Projects’ multisite trial contrasting standard evidence-based treatments, modular treatment, or usual care. We included an ethnically diverse sample of 46 youth (ages 7–13) who received treatment for disruptive behavior in modular treatment or usual care. A reliable observational coding system was developed to assess therapists’ in-session use of psychoeducation strategies (e.g., discussing causes of misbehavior, describing and providing rationale for treatment, etc.), as well as other engagement strategies (e.g., collaborative goal setting, managing expectations, etc.), in the early phase of treatment. Findings revealed that modular treatment therapists provided more psychoeducation and other engagement strategies compared with usual care therapists. Furthermore, psychoeducation strategies employed by therapists early on uniquely predicted subsequent parent involvement in treatment, over and above the use of other engagement strategies. Finally, therapists’ use of the psychoeducation strategy of discussing causes of child’s misbehavior mediated the effect of treatment condition on parent involvement in their child’s therapy. These ﬁndings suggest that the implementation of psychoeducation strategies upon entry into care promotes parent involvement in child psychotherapy for disruptive behavior.

During the time of this study, the Research Network on Youth Mental Health included Bruce F. Chorpita, Ann Garland, Robert Gibbons, Charles Glisson, Evelyn Polk Green, Kimberly Hoagwood, Kelly Kelleher, John Landsverk, Stephen Mayberg, Jeanne Miranda, Lawrence A. Palinkas, Sonja K. Schoenwald, and John R. Weisz (Network Director). The Modular Approach to Treatment of Children With Anxiety, Depression, or Conduct Problems (MATCH) manual used in this study was a precursor to a revised and expanded version for which Bruce F. Chorpita and John R. Weisz receive income. Correspondence should be addressed to Jonathan I. Martinez, Department of Psychology, California State University, Northridge, 18111 Nordhoff Street, Northridge, CA 91330-8255. E-mail: [email protected] 2 MARTINEZ ET AL.

In recent decades, there have been important advances Kim, & Quintana, 2002). Parents are more likely to drop in the child mental health (MH) care evidence base, out of services if they have a mismatch of treatment including an understanding of the etiology of child expectations with the therapist, such as how much MH problems, as well as the proliferation of evidence- parents need to be involved in their child’s treatment based interventions for those problems (National (Flisher et al., 1997; Nock, Ferriter, & Holmberg, 2007; Research Council and Institute of Medicine, 2009). Nock & Kazdin, 2005). One study found that parents’ Randomized controlled trials demonstrate that child perception of treatment relevance for their child’s prob- therapies are efficacious (National Research Council lems was the factor that best discriminated treatment and Institute of Medicine, 2009; Weisz & Kazdin, dropouts from completers (Morrissey-Kane & Prinz, 2010). However, as many as 75–80% of U.S. youth 1999). For ethnic minority and socially disadvantaged in need do not receive MH services (MHS), with the families, unpacking gaps in MH literacy may be critical disparity in need and service use highest among racial= for engaging parents. For example, addressing parent ethnic minority and socially disadvantaged families beliefs about MHS at treatment outset was associated (Kataoka, Zhang, & Wells, 2002). Among families with increased parent engagement in care in a sample who initiate services, more than 50% fail to complete of low-income, predominantly ethnic minority youth treatment (Kazdin, 1996). In fact, the majority of and their families (McKay, Nudelman, McCadam, & families who enter care attend only a small number of Gonzales, 1996; McKay, Stoewe, McCadam, & Gonzales, sessions insufficient to receive an adequate dose of treat- 1998). A qualitative study with Mexican American ment (Armbruster & Fallon, 1994; Harpaz-Rotem, parents found that most parents initiating child MHS Leslie, & Rosencheck, 2004). Thus, despite significant experienced disapproval from family who felt there was advances in developing evidenced-based interventions no MH problem or were skeptical of treatment and for child MH problems, they are limited in impact due expressed that Latinos were less likely to receive MHS to a lack of family engagement in services. because of a lack of knowledge about effective treat- In child psychotherapy, parents’ active involvement ments. In addition, parents were more likely to drop in treatment is considered critical to achieving successful out if they endorsed unrealistic beliefs and expectations outcomes (Nock & Kazdin, 2005), with children faring for child therapy (McCabe, 2002). better when parents are actively involved relative to In child psychotherapy, presenting factual infor- individual child treatment (Dowell & Ogles, 2010; mation to parents about their child’s MH problems

Karver, Handelsman, Fields, & Bickman, 2006). One and effective treatments is commonly referred to as of the most common presenting problems in children psychoeducation (Lukens & McFarlane, 2004). Hoagwood is disruptive behavior (Garland et al., 2001). Many et al. (2010) identified components of effective first-line interventions for disruptive behavior problems psychoeducation-based programs, with components necessitate active parent engagement (e.g., parent including (a) education on the nature of child MH prob- management training; Hoagwood et al., 2010; Weisz & lems, (b) discussion of family factors impacting child Kazdin, 2010). Parent engagement is defined as con- MH, (c) clarification of the child’s need for services, sistent treatment attendance, active participation in and (d) education on the nature of treatment. Psycho- sessions, and adherence to treatment recommendations education has emerged as an evidence-based practice between sessions (Nock & Ferriter, 2005). However, across a variety of child MH interventions (Fristad, many families seeking services often encounter a range 2006; Lukens & McFarlane, 2004) and may be a power- of barriers to persistence in MH care, including both ful tool for addressing perceptual barriers to engaging structural barriers (e.g., access to providers, financial families, particularly families with low MH literacy. As issues, transportation problems) and perceptual barriers psychoeducation facilitates the comprehension of com- about MH problems and treatments (e.g., recognition of plex information, families receive optimistic messages MH needs, knowledge about treatments, stigma related about the treatability of child MH problems (Miklowitz to care; Owens et al., 2002). Knowledge and beliefs about & Goldstein, 1997). Parents are treated as partners in the nature of MH problems and effective treatments— treatment based on the assumption that the more referred to as MH literacy—appear to be among the knowledgeable the parent, the greater the chance of most common and explanatory barriers, and may limit positive child outcomes (Lukens & McFarlane, 2004). the use of evidence-based interventions (Jorm, 2000). Orienting and preparing families for treatment using Addressing MH literacy concerns at treatment psychoeducation strategies may help alter misconcep- entry may facilitate the engagement of parents in child tions that derail engagement. therapy, as discrepancies in beliefs between parent and A range of interventions, in addition to psychoeduca- provider may shape differing treatment expectations tion, have been developed to improve engagement of and goals, which in turn jeopardizes engagement families in care. These other engagement interven- (Armbruster & Fallon, 1994; McKay, Harrison, Gonzales, tions have demonstrated increased attendance at initial PSYCHOEDUCATION AND PARENT ENGAGEMENT 3 intake (McKay et al., 1996; Szapocznik et al., 1988), problems. As engagement interventions have demon- throughout treatment (Chacko et al., 2009; McKay strated increased family engagement in the literature, et al., 1998; Nock & Kazdin, 2005; Prinz & Miller, we were particularly interested in examining the unique 1994), and greater treatment motivation and adherence effect of psychoeducation strategies on promoting parent (Chacko et al., 2009; Nock & Kazdin, 2005). Many engagement. To investigate this aim, an observational such engagement approaches emphasize a collaborative, coding system was developed to code treatment sessions problem-solving component to address practical barriers recordings for therapist in-session use of psychoeduca- (Chacko et al., 2009; McKay et al., 1996; Nock & Kazdin, tion and other engagement strategies. Data for this study 2005; Szapocznik et al., 1988). Several are based on were drawn from a community-based trial that included motivational interviewing (Miller & Rollnick, 1991), therapists randomized to provide an evidence-based which assume active resistance or lack of readiness to treatment approach (modular treatment [MT]) or usual change and thus address motivational barriers (Dishion care (UC) for children with disruptive behavior. & Kavanagh, 2000; Nock & Kazdin, 2005; Szapocznik The MT protocol featured modules based on common et al., 1988). Relatedly, some protocols ensure that elements of evidence-based treatments for youth with concerns from previous negative experiences in care are disruptive behavior (e.g., praise, rewards, psychoeduca- addressed (Dishion & Kavanagh, 2000; McKay et al., tion, etc.), which allowed therapists to flexibly deliver 1996). Although general treatment expectations are these evidence-based practices to their clients over the targeted in some form, these engagement interventions course of treatment. may not address parents’ attributions about their child’s Given the evidence that therapist use of psycho- problems, or provide information on problem etiology or education strategies is associated with increased parent rationale for treatment. Thus, there appears to be engagement (Hypothesis 1), we predicted a unique a relative emphasis on motivational and practical barriers effect of therapist use of psychoeducation strategies on and somewhat less attention to MH literacy barriers and promoting parent engagement while controlling for use of psychoeducation strategies for engaging families. therapist use of other engagement strategies. Because However, as discussed previously, engagement the MT protocol permitted flexible use of a psychoedu- strategies that go beyond addressing practical barriers cation module, along with evidence that UC therapists and treatment motivation are particularly indicated rarely deploy psychoeducation strategies with sufficient for families with low levels of MH literacy. For these intensity to promote a well-developed understanding families, psychoeducation can target problem recognition of child MH problems and treatments (Garland et al., barriers by increasing parents’ understanding of child 2010; Hypothesis 2) we predicted there would be more MH problems and addressing misperceptions of child observable therapist use and extensiveness of psychoe- MH needs. Psychoeducation can also target mispercep- ducation in MT versus UC. Last, (Hypothesis 3) we tions about care, where the rationale for evidence-based predicted that the higher levels of parent engagement interventions can be provided. Indeed, psychoeducation in MT versus UC would be explained by therapist use strategies have resulted in increased knowledge about of psychoeducation strategies while controlling for child MH problems (Brent, Poling, McKain, & Baugher, therapist use of other engagement strategies. 1993;Fristad,2006), increased treatment attendance (Becker et al., 2013;Fristad,2006), adherence with METHOD treatment recommendations (Pavuluri et al., 2004), and cognitive preparation of families entering services (Becker Data Source et al., 2013). Although there is evidence that therapists in usual care settings provide psychoeducation strategies The current study utilized data from the Child System to parents in the treatment of child disruptive behavior, and Treatment Enhancement Projects’ (Research Net- they rarely deploy these strategies with sufficient work on Youth Mental Health) multisite trial (Weisz intensity to promote a well-developed understanding of et al., 2012), a randomized effectiveness trial that tested child MH problems and treatments (Garland et al., 2010). standard and modular arrangements of evidence-based treatments compared with UC procedures in community clinic settings in Hawaii and Massachusetts. The overall THE CURRENT STUDY trial included 84 therapists providing treatment to 174 clinically referred youth (ages 7–13) for problems related The central aim of the current study was to examine to anxiety, depression, and=or disruptive behavior. the unique effect of therapist use of psychoeducation Therapists were randomly assigned to provide child strategies, over and above the use of other engagement treatment in one of three conditions: (a) UC procedures strategies, on promoting parent engagement in treatment in their clinics, or with evidence-based practices among families of children with disruptive behavior deployed in two forms; (b) standard manual treatment, 4 MARTINEZ ET AL. using full treatment manuals as they have been tested TABLE 1 in previous clinical trials; or (c) MT, in which therapists Sample Descriptives by Treatment Condition learn the component practices of all of the standard Total Sample manuals but individualize the use of the components MT (n ¼ 25) UC (n ¼ 21) (N ¼ 46) for each child using a guiding clinical algorithm. n(%) n(%) n(%) Therapists in the MT condition used the Modular Youth Gender Approach to Therapy for Children (Chorpita & Weisz, Male 21 (84.0) 14 (66.7) 35 (76.1) 2005), a collection of modules designed to correspond Female 4 (16.0) 7 (33.3) 11 (23.9) to evidence-based treatment procedures. M Youth Age (SD) 9.6 (2.0) 10.4 (1.7) 10.0 (1.9) All 84 therapists from the multisite trial were Youth Race=Ethnicity potentially eligible for the current study, 80% of which Non-Hispanic White 12 (50) 7 (33.3) 19 (42.2) Africa American 3 (12.5) 4 (19.0) 7 (15.6) were women, with a mean age of 40.6 years, and self- Latino 2 (8.3) 1 (4.8) 3 (6.7%) identified as the following: 56% White, 23% Asian Mixed 6 (25) 8 (38.1) 14 (31.1%) American, 6% African American, and 6% Pacific Islander. Other 1 (4.2) 1 (4.8) 2 (4.4) The mean number of years of clinical experience was 7.6 Note. One participant chose not to identify their race=ethnicity. years, and 40% were social workers, 24% were psycho- MT ¼ modular treatment; UC ¼ usual care. logists, and 36% were classified as ‘‘other.’’ Therapists reported the following orientations: 38% cognitive behavioral, 23% eclectic, 15% psychodynamic, 8% behavioral, 8% family systems, and 8% other. There were Measures 174 youth participants 7 through 13 years of age, with Psychoeducation and engagement strategies a mean age of 10.59 years (SD ¼ 1.76), 70% of which observational coding system (Martinez, Lau, & Chorpita, were boys; participants included 45.0% Caucasian, 32% 2012). We developed an observational coding system multiethnic, 9% African American, 6% Latino=a, 4% to code therapy session recordings from the early Asian American or Pacific Islander, 2% ‘‘other’’, and phase of treatment (i.e., first three treatment sessions) 2% who chose to not identify their ethnicity. Annual to measure therapists’ in-session use of psychoeducation family income was less than $40,000 for 55% of the and other engagement strategies directed at parents in sample, $40,000–$79,000 for 28% of the sample, $80,000– child therapy sessions. The coding system was developed $119,000 for 12% of the sample, and $120,000 or to align with the structure of the Therapy Process more for 6% of the sample (see Weisz et al., 2012,for Observational Coding System–Strategies Scale (McLeod a complete description of the overall trial’s study design). & Weisz, 2010) to characterize psychoeducation and other engagement strategies common in the child therapy Participants evidence base. The coding system follows the structure We only included children who were treated for disruptive of the Therapy Process Observational Coding System– behavior, as this treatment evidence base demonstrates Strategies Scale in that it includes both Microanalytic that active parent participation is indicated and integral and Extensiveness Scales. The Microanalytic Scale to the delivery of first-line interventions (e.g., parent is intended to track the occurrence and frequency management training), more so than anxiety and of therapist use of specific psychoeducation and other depression child treatments. Treatment within MT and engagement strategies over the course of a session. The UC arms were assessed, as these treatments offer flexi- Extensiveness Scale is designed to capture the extent to bility in the dose (amount and intensity) of therapist use which the therapist follows-through with each psycho- of psychoeducation and other engagement strategies. education and engagement strategy, with each strategy Thus, of the 184 youth participants in the Child System rated at the end of the session on a 7-point Likert scale. and Treatment Enhancement Projects Clinic Treatment The psychoeducation codes consist of the following Project effectiveness trial, the current study subsample five strategies used by therapists: (a) describing child included participants who were treated in the MT or behavior problems, (b) discussing causes of child’s UC arms (n ¼ 135), received treatment for disruptive misbehavior, (c) describing goals of treatment, (d) behavior problems within these arms (n ¼ 66), had at least providing rationale for treatment, and (e) providing one treatment session after study randomization (n ¼ 59), strategies for managing misbehavior. The engagement and had at least one available therapy session recording in codes consist of the following six strategies used by the early phase of treatment (i.e., first three treatment ses- therapists: (a) collaborative goal setting, (b) validating sions; n ¼ 46). Of these 46 cases, n ¼ 25 and n ¼ 21 cases and affirming parent’s commitment to treatment, were treated in the MT and UC arms, respectively. (c) checking in about past experiences and addressing See Table 1 for subsample descriptives. concerns in treatment, (d) managing expectations and PSYCHOEDUCATION AND PARENT ENGAGEMENT 5 what can work, (e) defining roles in treatment process, TABLE 3 and (f) addressing and problem-solving barriers to Interitem Correlations of Psychoeducation Strategies treatment (see Table 2 for coding definitions of psycho- Variable 1 2 3 4 5 education and engagement codes). The MT protocol featured psychoeducation and engagement modules 1. Describing Child Behavior Problems — that would allow MT therapists to flexibly deliver 2. Discussing Causes of Child’s .63 — Misbehavior psychoeducation and other engagement strategies 3. Describing Goals of Treatment .33 .06 — throughout the course of treatment. As the MT protocol 4. Providing Rational for Treatment .13 .02 .31 — included these modules that were derived from the child 5. Providing Strategies to Manage .25 .32 .40 .46 — therapy evidence base, some of the codes were Misbehavior based from the MT protocol. In addition to these codes, psychoeducation and engagement composite scores were derived by summing the individual psycho- to the low base rate of individual engagement strategies education and engagement strategies from the 7-point observed in session recordings, only the engagement Likert scale. Thus, the psychoeducation composite composite score was used to capture therapists’ overall comprised the sum of the five individual psycho- use of other engagement strategies. education strategies (composite total out of 35), and the internal consistency of this composite was in the Scoring strategy. The scoring strategy includes acceptable range (a ¼ 0.78; see Table 3 for interitem a Microanalytic Scale for raters to code the occurrence correlations of psychoeducation items). The engagement of specific psychoeducation or other engagement strategies composite comprised the sum of the six individual during a given time segment (defined as 5-min periods) engagement strategies (composite total out of 42). Due with an entire treatment session. The Microanalytic Scale

TABLE 2 Psychoeducation and Engagement Coding Deﬁnitions

Codes Deﬁnition

Psychoeducation Codes

Describing Child Behavior Problems Extent to which the therapist describes child behavior problems in general (e.g., symptoms, diagnosis, impairment) and behavior problems specific to the child. Discussing Causes of Child’s Misbehavior Extent to which the therapist discusses general factors that may contribute to child misbehavior, elicits specific factors from parent, and reflects=summarizes contributing factors. Describing Goals of Treatment Extent to which the therapist describes what will occur during treatment sessions and describes specific treatment goals or changes that can occur for parent and child. Providing Rationale for Treatment Extent to which the therapist provides evidence of efficacy of treatment, discusses consequences of untreated behavior problems, and provides informational handouts about treatments for problems. Providing Strategies to Manage Misbehavior Extent to which therapist provides strategies to manage child’s misbehavior and provides rationale for strategies. Engagement Codes Collaborative Goal Setting Extent to which therapist elicits main challenges going on with child or family, elicits goals or changes parent would like to see occur in child, and reviews=clarifies main challenges or goals. Validating & Affirming Parent Commitment to Treatment Extent to which therapist validates parent as caring adult, emphasizes that parent is expert on child, and reminds parent of his=her invaluable role in treatment process. Checking in About Past Experiences & Addressing Extent to which therapist elicits from parent what has worked well= Concerns in Treatment positive experiences and not worked well=negative experiences in treatment, and elicits concerns parent has with current treatment. Managing Expectations and What Can Work Extent to which therapist describes what will or will not occur in treatment process, and emphasizes working with parent to provide strategies to manage child’s behavior. Defining Roles in Treatment Process Extent to which therapist describes parent, child, and therapist role in treatment and reinforces participation. Addressing and Problem-Solving Barriers to Treatment Extent to which therapist elicits from parent potential barriers to parent=child participation, and helps problem solve barriers. 6 MARTINEZ ET AL. yields occurrence and observed frequency of use of each which were used to refine the coding system items. After therapy practice. The scoring strategy also includes the piloting phase was completed, a final version of the Extensiveness ratings to measure the extent to which coding manual was produced and utilized throughout therapists deployed psychoeducation and other engage- the current study. ment strategies within a treatment session on a 7-point Likert scale, ranging 1 (not at all), 3 (somewhat), 5 Therapy session recordings. For the 46 cases in (considerably), and 7 (extensively). Extensiveness ratings the current sample, all therapy session recordings in reflect both frequency and thoroughness of use of the early phase of treatment (i.e., first three treatment the strategy. Thoroughness is determined by the detail sessions) were coded. A total of 105 available session and=or quality in which the therapist employs the recordings out of 138 possible session recordings strategy, and=or the extent to which the therapist (76.1%) in the early the treatment phase were available, follows-through with the strategy (i.e., covering different with a per case average of 2.28 recordings (SD ¼ .83). Of components of the strategy in full and not abandoning these 105 session recordings, 64 included sessions with the strategy after limited use). A low Extensiveness parents (61.0%). As the observational coding system rating reflects a cursory and=or incomplete application measured therapist use of strategies directed at parents of the treatment strategy with limited follow-through, in child therapy sessions, sessions that did not include whereas a high Extensiveness rating reflects a high a parent received the lowest rating for all psychoeduca- degree of effort or force that the therapist places in tion and engagement codes. delivering the treatment strategy. Definitions for each anchor point on the 1-to-7 Likert scale relevant to each strategy were provided in the coding manual. Parent Engagement Measures Parent involvement. Sessions in which a parent was Coders. The coding team consisted of the first present at their child’s therapy session were documented author and four undergraduate research assistants for each case from therapy session recordings. Each under the supervision of a clinical psychologist therapy session recording captured whether the session (coauthor). Four coders were trained for 8 consecutive included the child client only, the parent only, or both weeks before being released to code recordings indepen- the parent and child. The proportion of sessions in dently. Training of coders included didactic training on which a parent was present throughout their child’s the coding manual, practice scoring of sessions, review treatment episode (i.e., the number of parent only plus of specific session segments, and weekly coding meetings parent and child sessions out of the total number of to clarify codes and arrive at consensus. Interrater therapy sessions) was documented and used as the reliability was calculated across all coders for each of outcome variable of parent involvement. Therapist use the items using intraclass correlations (ICCs). Given of psychoeducation and other engagement strategies our coding design where we trained a group of four measured in early treatment (i.e., first three treatment coders who were each randomly assigned different sessions) were used as a predictors of parent involve- recordings to code, we used a two-way mixed average ment beyond the initial phase of treatment (i.e., beyond measures with absolute agreement ICC. Coders were the third treatment session). Parent involvement beyond approved for coding independently after their ratings the third session was used as the key outcome variable, achieved acceptable interrater reliability (ICC .60; as including early coded treatment sessions would Cicchetti, 1994). Weekly coding meetings were held result in counting early session involvement toward the throughout the duration of coding project to prevent outcome variable that is being predicted. rater drift. Therapeutic alliance. The quality of the parents’ Pilot coding phase. A pilot coding phase was con- working alliance with their children’s therapists was ducted to aid in the development of items for the coding assessed at study completion (2-year follow-up) via the system as well as train coders to establish adequate pre- Therapeutic Alliance Scale for Children (Shirk & Saiz, study interrater reliability. A random sample of session 1992). The nine-item parent-report form has shown tapes from early treatment sessions in the UC and MT good internal consistency (a ¼ .92) and good 7-day to arms were coded. Based on training in a detailed coding 14-day test–retest reliability (r ¼ .82) in samples of manual, four undergraduate student coders indepen- parents of clinic-referred youth. In the current sample, dently coded five therapy sessions. Items that demon- the parent–therapist alliance scale demonstrated good strated poor agreement during the pilot phase were internal consistency (a ¼ .84). Sample items included refined or dropped. During the piloting phase, coders ‘‘I looked forward to meeting with my child’s therapist, provided feedback on item content and definitions, I liked spending time with my child’s therapist, and I feel PSYCHOEDUCATION AND PARENT ENGAGEMENT 7 like my child’s therapist was on my side and tried to help variable). In the first step, the engagement composite me.’’ Response items were on a 4-point Likert scale: 1 was entered into each model. In the second step, the (not like me), 2 (a little like me), 3 (mostly like me), psychoeducation strategies were entered into the model. and4(very much like me). To determine whether multicollinearity among psychoeducation predictors impacted coefficient estimates due to the potential of unstable standard errors, Tolerance Satisfaction with services. The Client Satisfaction and the Variance Inflation Factor (VIF) were examined. Questionnaire–Parent Report (CSQ-8=P; Attkisson & Tolerance is an indication of the percentage of variance Greenfield, 2004) consists of an eight-item parent-report in the predictor that cannot be accounted for by the on satisfaction with their child’s services, which was other predictors, with values less than .10 indicating that assessed at study completion (2-year follow-up). The a predictor is redundant, and VIF (1=tolerance) values CSQ-8 has demonstrated good internal consistency in greater than 10 suggesting a high degree of multicol- numerous studies with diverse client samples (a ranges ¼ linearity (Chen, Ender, Mitchell, & Wells, 2003). These .83–.93). In the current sample, the CSQ-8=P demon- multicollinearity diagnostics revealed that Tolerance strated good internal consistency (a ¼ .91). Sample items and VIF values for each of the five psychoeduction included ‘‘How would you rate the quality of services your predictors were in the acceptable range, indicating that child received, Did you get the kind of service you wanted using these psychoeducation items in a single model is for your child, How satisfied are you with the amount justified. Three additional hierarchical models were of help your child received?’’ Response items were on tested in a similar fashion using the psychoeducation a 4-point Likert scale: 1 (quite dissatisfied), 2 (indifferent composite replacing the individual psychoeducation or mildly dissatisfied), 3 (mostly satisfied), and 4 (very strategies, thus producing a total of six hierarchical satisfied). regression models. Table 4 displays the results of each step of the six hierarchical models. RESULTS In the first step for all hierarchical regression models, therapist use of other engagement strategies Reliability of Psychoeducation and Engagement (engagement composite) in the initial phase of treatment Strategies Observational Coding System was significantly associated with subsequent parent involvement (B ¼ .04, p < .05) but was unrelated to other

Of the 64 coded sessions recordings in which a parent outcome variables (alliance and satisfaction). For the was present, 15 (23.4%) were randomly selected for three models using the individual psychoeducation double-coding to examine interrater reliability using strategies, the results are as follows. In the second step, two-way mixed average measures with absolute therapist use of psychoeducation strategies in the agreement intraclass correlations (ICCs). ICCs below initial phase of treatment was found to be uniquely .40 reflect ‘‘poor’’ agreement, .40–.59 reflect ‘‘fair’’ associated with subsequent parent involvement (two of agreement, .60–.74 reflect ‘‘good’’ agreement, and .75 five psychoeducation predictors significant: Discussing and above reflect ‘‘excellent’’ agreement (Cicchetti, causes of misbehavior B ¼ .13, p < .05; Describing goals 1994). The average ICCs for the five psychoeducation of treatment B ¼ .14, p < .05) while controlling for Extensiveness codes and the Microanalytic occurrence therapist use of other engagement strategies. In codes were .78 and .75, respectively. The ICCs for the addition, three of five psychoeducation strategies were psychoeducation and engagement Extensiveness Scale significantly associated with parent–therapist alliance; composites (total scores) were .85 and .66, respectively. however, these associations were inconsistent. Neither The ICCs for the psychoeducation and engagement therapist use of psychoeducation nor other engagement Microanalytic Scale composites (any occurrence) were strategies were related to parent satisfaction. .96 and .64, respectively. Thus, all codes demonstrated Three additional hierarchical models were tested acceptable reliability (ICCs >.60). using the psychoeducation composite in place of the five psychoeducation strategies for the second step. As predicted, therapists’ use of psychoeducation in the initial Effect of Psychoeducation on Promoting Parent phase of treatment was found to be uniquely associated Engagement with subsequent parent involvement (psychoeducation Three hierarchical regression models were tested to composite B ¼ .04, p < .05) while controlling for examine the unique effect of therapist use of psycho- therapist use of other engagement strategies. Of note, education strategies (predictor variables) on promoting therapist use of other engagement strategies was no parent engagement (three outcome variables: parent longer associated with parent involvement after involvement, alliance, and satisfaction) while accounting accounting for psychoeducation strategies. Therapist for therapist use of other engagement strategies (control use of psychoeducation (composite) was unrelated to 8 MARTINEZ ET AL.

TABLE 4 Hierarchical Regression Models of the Effects of Therapist Use of Psychoeducation and Other Engagement Strategies on Parent Engagement

Parent Involvement Model Alliance Model Satisfaction Model

Variable BSE b BSE b BSEb

Step 1 Control Variable Engagement Composite .04 .02 .35 .09 .16 .09 .07 .12 .10 Step 2a Psychoed Variables Engagement Composite .01 .02 .04 .12 .21 0.12 .08 .18 .11 Model Describing child behavior problems .07 .06 .22 2.53 .80 .89 .23 .56 .11 Discussing causes of child’s .13 .05 .49 2.04 .69 .79 .13 .46 .07 misbehavior Describing goals of treatment .14 .06 .38 .46 .66 .13 .04 .58 .04 Providing rationale for treatment .03 .10 .05 2.3 .96 .40 1.58 .86 .36 Providing strategies to manage .01 .06 .05 .13 .56 .05 .38 .50 .20 misbehavior Step 2a Psychoed Composite Engagement Composite .04 .02 .35 .02 .22 .02 .06 .17 .08 Model Psychoed Composite .04 .02 .50 .13 .18 .15 .02 .14 .03

aEach Step 2 model was tested hierarchically with Step 1, but not together. p < .05. p < .01. other outcome variables (alliance and satisfaction). The above 60%, with the exception of providing rationale psychoeducation composite models were also examined for treatment (27.7%). For the UC condition, the rate by treatment condition. Within the UC condition, thera- of occurrence of each psychoeducation strategy was pist use of psychoeducation was uniquely associated with below 40%. The psychoeducation composite occurred parent involvement (B ¼ .15, p < .01) while controlling in 95.7% of MT sessions, but only in 52.9% of UC for therapist use of other engagement strategies, which sessions. The engagement composite occurred in 87.0% was nonsignifcant (B ¼.05, p ¼ .41). Within the MT of MT sessions, but only in 58.8% of UC sessions. condition, therapist use of psychoeducation was not asso- Independent samples t tests were conducted to exam- ciated with parent involvement (B ¼ .02, p ¼ .46) while ine differences in psychoeducation strategies and other controlling for therapist use of other engagement strate- engagement strategies delivered to parents in MT and gies, which was also nonsignifcant (B ¼.02, p ¼ .51). UC conditions (see Table 5). For these analyses, to obtain scores for each child case (as opposed to scores for each session recording), Extensiveness ratings for Psychoeducation and Other Engagement early treatment sessions were averaged (i.e., mean Strategies Observed in MT Versus UC Sessions Extensiveness ratings for Sessions 1–3 for each child For early treatment sessions in which a parent was case). Compared with UC therapists, MT therapists present, the occurrence rates for therapist use of received signiﬁcantly higher Extensiveness ratings for psychoeducation and other engagement strategies were the delivery of the individual psychoeducation strategies examined by condition. For the MT condition, the rate and composite, as well as the engagement composite. of occurrence of each psychoeducation strategy was Cohen’s d was assessed as a measure of effect size

TABLE 5 Average Psychoeducation Extensiveness Ratings (Case Level) by Treatment Condition

Psychoeducation Strategies MTa UCb t(44) ES

Describing Child Behavior Problems 2.8 (1.7) 1.2 (0.5) 4.6 1.3 Discussing Causes of Child’s Misbehavior 3.2 (1.8) 1.3 (0.9) 4.5 1.3 Describing Goals of Treatment 2.7 (1.4) 1.3 (0.6) 4.6 1.3 Providing Rational for Treatment 1.5 (0.9) 1.1 (0.3) 2.5 0.6 Providing Strategies to Manage Misbehavior 3.2 (1.7) 1.2 (0.7) 5.3 1.5 Psychoeducation Composite 13.4 (4.6) 6.0 (2.3) 7.1 2.0 Engagement Composite 12.7 (3.9) 7.2 (1.9) 6.3 1.8

Note. MT ¼ modular treatment; UC ¼ usual care; ES ¼ effect size ¼ Cohen’s d, which is the difference between the means of the two study groups divided by the pooled standard deviation. an ¼ 25. bn ¼ 21. p < .05. PSYCHOEDUCATION AND PARENT ENGAGEMENT 9 for MT versus UC differences in deployment of mediational effect of psychoeducation. Estimates of all psychotherapeutic strategies, which almost all differences paths are calculated using ordinary least squares regression. demonstrating large effects (Cohen’s d > 0.8). The Preacher and Hayes (2008) approach is the preferred method for testing mediation in the current study, in lieu of the traditional Baron and Kenny Parent Engagement in Child Therapy by Treatment (1986) approach, as it allows testing for multiple media- Condition (MT Versus UC) tors in one single model. Preacher and Hayes delineated several advantages to testing a single multiple mediation Participants attended an average of 14.59 (SD ¼ 10.74) model instead of running separate simple mediation sessions (range ¼ 2–41) during the study period. Of the models. First, testing multiple mediators allows one to 105 coded early treatment sessions (i.e., Sessions 1–3), conclude whether a set of mediator variables mediates parents were involved (i.e., present) in 64 sessions the effect of X on Y. Second, it is possible to determine (61.0%). As predicted, parent involvement varied signifi- to what extent potential variables mediate the effect of X cantly by treatment condition, with MT parents being on Y, conditional on the presence of other mediators in involved in more sessions than UC parents. For early the model (in essence, controlling for other mediators treatment sessions, UC parents were involved in 17 in the model). Third, when multiple mediators are out of 53 sessions (32.1%), whereas MT parents were tested in a multiple mediation model, the likelihood of involved in 47 of 52 sessions (90.4%; t ¼ 7.56, parameter bias due to omitted variables is reduced, as p < .001). Parent involvement beyond the early phase opposed to running several separate mediation models of treatment (i.e., after third session) also varied signifi- that may lead to biased parameter estimates (due to cantly by treatment condition, with UC parents being omitted variables). Fourth, including several mediators involved in 28.0% of sessions and MT parents being in one model allows one to determine the relative involved in 71.6% of sessions (t ¼ 3.6, p < .01). There magnitude of the specific indirect effects associated were no significant differences between MT and UC with each mediator variable. Fifth, this method uses conditions in parent-reported measures of satisfaction Bootstrapping, a nonparametric resampling procedure with child MHS and parent–therapist alliance, and in which indirect effect estimates are calculated across therefore mediation analyses focused on parent involve- 5,000 bootstrap samples, along with 95% confidence ment as the engagement outcome of interest. intervals for the indirect effects. Bootstrapping is

a preferred test for determining significant mediation (i.e., significant indirect effect) over the Sobel test, as it Psychoeducation as a Mediator of Treatment makes no sampling distributional assumptions of the Approach on Parent Engagement indirect effect, and thus can be effectively utilized with The proposed mediational relationship was examined smaller sample sizes while preserving power to detect a using the INDIRECT macro for SPSS (Preacher & significant indirect effect. Given these advantages, this Hayes, 2008), which estimates the total, direct, and analytic approach is preferred for the current study indirect effects of a predictor variable on an outcome given the small sample size, testing for multiple psychoe- variable through a proposed mediator variable (or mul- ducation mediators simultaneously, and adjusting all tiple mediators) and allows controlling for the influence paths for the potential influence of the engagement cov- of other variables. The total effect is the effect of X on Y ariate. in the absence of the Mediator (known as path C); for Figure 1 depicts findings of the mediation model. the current study, this is the effect of treatment con- This mediation model used the psychoeducation com- dition on parent involvement in the absence of therapist posite as the hypothesized mediator and controlled for use of psychoeducation. Conversely, the direct effect therapists’ use of other engagement strategies. The effect is the effect of X on Y while controlling for the Mediator of treatment condition (MT vs. UC) on psychoeduca- (known as path C’); for the current study, this is the tion (hypothesized mediator) was significant (B ¼ 4.56, effect of effect of treatment condition on parent p < .01). The direct effect of psychoeducation on parent involvement while controlling for therapist use of psycho- involvement was not significant (B ¼ .03, p ¼ .08). How- education strategies. The indirect effect is a measure off ever, this path is not required to be significant to test for the amount of mediation and represents the portion mediation using the Preacher and Hayes (2008) method. of the relationship between X and Y that is mediated The total effect of treatment condition on parent by the proposed Mediator; for the current study, this is involvement was significant, with greater parent involve- portion of the relationship between treatment condition ment in MT versus UC (B ¼ .40, p < .05). However, the and parent involvement that is mediated by therapist direct effect of treatment condition on parent involve- use of psychoeducation strategies. Therefore, the indirect ment was not significant (B ¼ .25, p ¼ .17). That is, the effect was assessed as a measure of the hypothesized effect of treatment condition on parent involvement 10 MARTINEZ ET AL.

FIGURE 1 Mediation model, controlling for therapist use of other engagement strategies. Note. MT ¼ modular treatment; UC ¼ usual care. was no longer significant when controlling for psychoe- mediator model). There was a significant effect of treat- ducation, suggesting the presence of mediation. To for- ment condition on describing child behavior problems mally test for mediation (i.e., a significant indirect effect (B ¼ 1.15, p < .05) and discussing causes of child’s of treatment condition on parent involvement via the misbehavior (B ¼ 1.61, p < .01). The direct effect of these hypothesized mediator of psychoeducation), the boot- hypothesized mediators on parent involvement was only strap confidence intervals were assessed, which revealed significant for discussing causes of child’s misbehavior no significant indirect effect (Bootstrap M of indirect (B ¼ .12, p < .05). The total effect of treatment condition effect B ¼ .15), 95% confidence interval [.02, .40]. on parent involvement was significant, with greater Thus, the hypothesized mediational effect of therapist parent involvement in MT versus UC (B ¼ .40, p < .05). use of psychoedcation on the relationship between treat- However, the direct effect of treatment condition on ment condition and parent involvement did not reach parent involvement was not significant (B ¼ .22, statistical significance. p ¼ .23). That is, the effect of treatment condition The INDIRECT command allows testing for mul- on parent involvement was no longer significant when tiple mediators in one model, and therefore the individ- controlling for multiple psychoeducation mediators, ual psychoeducation strategies were entered as multiple suggesting the presence of mediation. To formally test mediators while still controlling for therapist use of for mediation (i.e., significant indirect effects of treat- other engagement strategies (see Figure 2 for multiple ment condition on parent involvement via specific

FIGURE 2 Multiple mediation model, controlling for therapist use of other engagement strategies. Note. MT ¼ modular treatment; UC ¼ usual care. PSYCHOEDUCATION AND PARENT ENGAGEMENT 11 psychoeducation mediators), the bootstrap confidence the focus of treatment have been associated with intervals were assessed, and revealed a significant improved family attendance (Brookman-Frazee, Haine, indirect effect for discussing causes of child’s misbehavior Gabayan, & Garland, 2008; Nock & Kazdin, 2001). (Bootstrap M of indirect effect B ¼ .19), 95% confidence Preparing and orienting parents to their child’s treat- interval [.001, .52]. Thus, the hypothesized mediational ment using psychoeducation strategies can help address effect of therapist use of psychoeducation strategies on parent misconceptions about child MH problems and the relationship between treatment condition and parent unrealistic treatment expectations, thereby increasing involvement reached statistical significance for the parent involvement in services. Indeed, a recent study strategy of discussing causes of child’s misbehavior. found that engagement interventions that outperformed other engagement interventions on indicators of family engagement utilized psychoeducation as a core DISCUSSION element (Becker et al., 2013). As hypothesized, there was more observable therapist The current study examined the extent to which thera- use and extensiveness of psychoeducation and other pists used psychoeducation and other engagement stra- engagement practices in MT versus UC early treatment tegies, consistent with evidence-based practices, within sessions. The individual psychoeducation strategies and a randomized effectiveness trial for children treated with overall composite occurred in the majority of early evidence-based approaches versus usual care (UC). treatment MT sessions but infrequently in UC sessions. This study provides detailed data on the variability of When psychoeducation strategies were deployed by UC psychoeducation and other engagement strategies therapists, they were delivered with much lower deployed by therapists in evidence-based MT and in extensiveness compared to MT therapists. As the MT UC. We hypothesized that therapist use of psychoedu- protocol featured a psychoeducation module that would cation strategies in the early phase of treatment would allow therapists to flexibly use psychoeducation with uniquely promote subsequent parent engagement in their clients over the course of treatment, this finding child treatment for disruptive behavior, after accounting is not surprising. Yet this finding is also consistent for therapist use of other engagement strategies. This with research that suggests that directive treatment was partially supported by the study findings, as thera- approaches (such as psychoeducation) are not observed pist use of psychoeducation strategies uniquely pre- as frequently in UC compared to evidence-based treat- dicted parent involvement beyond the initial treatment ment models (Malik, Beutler, Alimohamed, Gallagher- phase while accounting for therapist use of engagement Thompson, & Thompson, 2003). Although UC strategies. However, therapist use of psychoeduation therapists have positive attitudes about psychotherapeu- strategies was inconsistently related to parent–therapist tic techniques that may be conceptualized as directive alliance and unrelated to parent satisfaction with their (Brookman-Frazee, Garland, Taylor, & Zoffness, 2009), child’s services. As a psychoeducation module was fea- when they use psychoeducation in their practice, they tured in the MT protocol that permitted flexible use of rarely deliver these strategies with sufficient intensity psychoeducation throughout the course of treatment, that would be consistent with the expectations of it was important to determine that our index of observed evidence-based treatment models (Garland et al., 2010). psychoeducation was not merely a proxy for adherence Similarly, other engagement strategies occurred in to the MT protocol. As such, it was particularly impor- the majority of early treatment MT sessions but tant to establish whether the association between psy- infrequently in UC sessions. When engagement strate- choeducation and parent engagement could be observed gies were deployed by UC therapists, they were delivered within the UC condition. This was evidenced, as UC at a much lower extensiveness than MT therapists. This therapists’ overall use of psychoeducation in early treat- result is surprising when one considers that UC thera- ment was significantly associated with later parent pists often spend much time using eclectic strategies involvement, whereas therapist overall use of engagement to engage clients, often at the expense of delivering strategies was unrelated to parent involvement. evidence-based, cognitive and behavioral strategies In assessing the role of specific psychoeducation stra- (McLeod & Weisz, 2005). It is important to note that tegies, therapist use of strategies including discussing the engagement strategies that were observationally causes of child’s behavior problems and describing goals coded corresponded to practices found in evidence- of treatment program were found to significantly predict based interventions that utilize enhanced engagement parent involvement while controlling for therapist use of strategies. UC therapists may spend much time on such other engagement strategies. These findings align with things as joining empathically with parents to engage observational studies of predictors of parent involve- and build rapport in ways that may not necessarily ment in child therapy. For example, parents’ expecta- position the parent to be an active agent in the child’s tions about treatment and parent–youth agreement on treatment. In addition, UC therapists may place an 12 MARTINEZ ET AL. emphasis on engagement but may direct most of their discussing causes of child’s misbehavior, thus supporting attention on engaging their child clients rather than the presence of mediation. This specific psychoeducation the parents. This is supported by the significantly higher strategy may have accounted for differences in parent proportion of child-only sessions in UC relative to MT. involvement between treatment conditions due to Levels of parent engagement in treatment for child increasing parents’ understanding of factors associated disruptive behavior were characterized across treatment with child misbehavior. Parents who hold biopsychoso- conditions, and it was hypothesized that parents assigned cial etiological beliefs about causes of child problems are to receive MT would demonstrate higher engagement more likely to use MHS to address those problems in services than parents in UC. This hypothesis was (Yeh et al., 2005). Discussing causes of child’s misbeha- partially supported. As predicted, parents in the MT vior may target parent misperceptions about causes of condition were involved in a much higher proportion of misbehavior and highlight causes that can be targeted sessions than parents in the UC condition; however, there in evidence-based treatment for child disruptive behavior were no differences in parent ratings of the parent– that leverages parent involvement. therapist alliance and satisfaction with services between The results of the current study should be interpreted MT and UC conditions. That is, parents’ feelings of in light of some study limitations. First, although we having a collaborative relationship with their child’s used multiple measures of parent engagement, we did therapist and their satisfaction with their child’s services not examine the quality of parent engagement in session were similar across MT and UC conditions. This finding (e.g., actively contributing to therapeutic discussions may be due to a lack of sensitivity in these measures to and activities, asking questions, etc.). Second, session detect differences because of a ceiling effect and demand recordings were not available for all early treatment characteristics (i.e., consumers wanting to provide sessions due to missing data, but there were no positive feedback), as ratings of alliance and satisfaction differences in missingness in MT versus UC. It is con- were generally very high. Studies support that parents ceivable that some psychoeducation and engagement often express high satisfaction with services, regardless strategies occurred in sessions that were not recorded, of whether those services result in improvement in their and thus not captured in early treatment sessions adding children (Garland, Aarons, Hawley, & Hough, 2003). error to our analysis. Third, the small sample size did This finding that parent-rated measures of alliance and not permit the examination of moderators of parent satisfaction were no different between UC and MT involvement, such as racial=ethnic and socioeconomic may help dispel some of the concerns that therapists background. Fourth, as the MT condition had more have regarding the use of evidence-based practices initial sessions attended by a parent, we observationally interfering with the development of therapeutic alliance coded more MT sessions relative to the UC condition, and contributing to poorer engagement of families and thus may have had more of an opportunity to (Nelson, Steele, & Mize, 2006). observe therapist use of psychoeducation and other Finally, it was hypothesized that effect of treatment engagement strategies in the MT condition. Fifth, treat- condition (MT vs. UC) on parent involvement would ment for child behavior problems typically includes be explained by therapist use of psychoeducation some degree of psychoeducation, and it may be that strategies while controlling for use of other engagement we were measuring therapist use of strategies common strategies. This was partially supported by study find- to all behavioral treatments. Last, we did not include ings. Treatment condition significantly predicted parent the standard manual condition in the current study involvement (with greater parent involvement in MT vs. due to a lack of resources for observationally coding UC), and the direct effect of treatment condition on additional recordings in this treatment arm, which parent involvement was no longer significant when would have elucidated the extent to which therapists accounting for overall psychoeducation strategies. How- deploy psychoeducation and other engagement strate- ever, the indirect effect of treatment condition on parent gies while adhering to a particular treatment manual. involvement via psychoeducation did not reach statisti- Despite these limitations, the current study has impor- cal significance, thus indicating no significant mediation. tant merits that warrant consideration. The develop- There was evidence that therapist use of individual ment of a reliable observational coding system to psychoeducation strategies mediated the relationship examine therapist in-session use of psychotherapeutic between treatment condition and parent involvement, strategies, inclusion of a relatively diverse group of part- as discussing causes of child’s misbehavior with parents icipants and providers who are generally representative in initial sessions significantly predicted later parent of other samples from community-based MH settings, involvement beyond the early phase of treatment. assessing multiple indicators of parent engagement, and A formal test for mediation showed a significant indirect use of randomized sample of therapists in a larger effect of treatment condition on parent involvement via effectiveness trial are notable strengths of the current study. PSYCHOEDUCATION AND PARENT ENGAGEMENT 13

CONCLUSION REFERENCES

Consistent with the existing literature, we found great Armbruster, P., & Fallon, T. (1994). Clinical, sociodemographic, and variability in parent involvement among families receiv- systems risk factors for attrition in a children’s mental health clinic. ing care in community-based MH services for child dis- American Journal of Orthopsychiatry, 64, 577–585. doi:10.1037= h0079571 ruptive behavior. The findings of low occurrence and Attkisson, C. C., & Greenfield, T. K. (2004). The UCSF client extensiveness of psychoeducation and other engagement satisfaction scales: I. The client satisfaction questionnaire-8. In strategies deployed by UC therapists reflect a cursory M. Maruish (Ed.), The use of psychological testing for treatment and=or incomplete application of these psychotherapeu- planning and outcome assessment (3rd ed., pp. 799–811). Mahwah, tic strategies with limited follow-through. These findings NJ: Erlbaum. Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator are consistent with research that indicates gaps in using variable distinction in social psychological research: Conceptual, evidence-based practices in community-based service set- strategic, and statistical considerations. Journal of Personality and tings. Although the observational coding measure used in Social Psychology, 51, 1173–1182. doi:10.1037=0022-3514.51.6.1173 the current study is not a measure of fidelity to a parti- Becker, K. D., Lee, B. R., Daleiden, E. L., Lindsey, M., Brandt, N. E., cular treatment, the low occurrence and extensiveness & Chorpita, B. F. (2013). The common elements of engagement in children’s mental health services: Which elements for which of psychoeducation and other engagement strategies sug- outcomes? Journal of Clinical Child & Adolescent Psychology, 44, gests that these evidence-based strategies are not being 30–43. doi:10.1080=15374416.2013.814543 delivered in UC as thoroughly as would be expected in Brent, D. A., Poling, K., McKain, B., & Baugher, M. (1993). A psy- an evidence-based model. Taken together, these findings choeducational program for families of affectively ill children and replicate results from similar observational coding studies adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 32, 770–774. doi:10.1097=00004583-199307000-00010 that show psychotherapeutic strategies conceptually con- Brookman-Frazee, L., Garland, A. F., Taylor, R., & Zoffness, R. sistent with evidence-based practices for children with (2009). Therapists’ attitudes towards psychotherapeutic strategies disruptive behavior are delivered with some frequency in community-based psychotherapy with children with disruptive in UC but are not delivered with sufficient intensity. behavior problems. Administration and Policy in Mental Health There is a substantial amount of research on client and Mental Health Services Research, 36, 1–12. doi:10.1007= s10488-008-0195-6 characteristics that predict engagement in care. An under- Brookman-Frazee, L., Haine, R. A., Gabayan, E. N., & Garland, A. F. standing about how specific in-session psychotherapeutic (2008). Predicting frequency of treatment visits in community-based strategies may influence parent involvement in child youth psychotherapy. Psychological Services, 5, 126–138. doi:10.1037= psychotherapy further elucidates the therapist’s role in pro- 1541-1559.5.2.126 moting parent involvement. These results support the need Chacko, A., Wymbs, B. T., Wymbs, F. A., Pelham, W. E., Swanger-Gagne, M. S., Girio, E., ... O’Connor, B. (2009). for training community-based therapists in implementing Enhancing traditional behavioral parent training for single mothers psychoeducation strategies early in care to promote parent of children with ADHD. Journal of Clinical Child & Adolescent involvement in child MHS. This may be especially impor- Psychology, 38, 206–218. doi:10.1080=15374410802698388 tant for families with lower levels of MH literacy, such as Chen, X., Ender, P., Mitchell, M., & Wells, C. (2003). Regression ethnic minority and socially disadvantaged families that with SPSS. Retrieved from http://www.ats.ucla.edu/stat/spss/ webbooks/reg/default.htm may experience perceptual barriers to engagement in child Chorpita, B. F., & Weisz, J. R. (2005). Modular approach to therapy for MH care. Future studies should investigate whether psy- children with anxiety, depression, or conduct problems. Honolulu: choeducation strategies promote MH literacy among eth- University of Hawaii at Manoa; Boston, MA: Judge Baker nically diverse parents initiating services, and whether this Children’s Center, Harvard Medical School. results in increased parent engagement in child MH care. Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6, 284–290. doi:10.1037= 1040-3590.6.4.284 ACKNOWLEDGMENTS Dishion, T. J., & Kavanagh, K. (2000). A multilevel approach to family-centered prevention in schools: Process and outcome. The authors would like to acknowledge the UCLA Addictive Behaviors, 25, 899–911. doi:10.1016=S0306-4603(00)00126-X Culture and Minority Mental Health Lab and the Child Dowell, K. A., & Ogles, B. M. (2010). The effects of parent partici- F.I.R.S.T. lab for their contribution to the observational pation on child psychotherapy outcome: A meta-analytic review. Journal of Clinical Child & Adolescent Psychology, 39, 151–162. coding of therapy session recordings. doi:10.1080=15374410903532585 Flisher, A. J., Kramer, R. A., Grosser, R. C., Alegria, M., Bird, H. R., Bourdon, K. H., ...Hoven, C. W. (1997). Correlates of unmet need FUNDING for mental health services by children and adolescents. Psychological Medicine, 27, 1145–1154. doi:10.1017=S0033291797005412 This study was supported by a Ford Foundation Fristad, M. A. (2006). Psychoeducational treatment for school-aged Fellowship from the National Research Council of the children with bipolar disorder. Development and Psychopathology, 18, 1289–1306. doi:10.1017=S0954579406060627 National Academies. 14 MARTINEZ ET AL.

Garland, A. F., Aarons, G. A., Hawley, K. M., & Hough, R. L. (2003). and their caregivers. Health & Social Work, 23, 9–15. doi:10.1093= Relationship of youth satisfaction with mental health services and hsw=23.1.9 changes in symptoms and functioning. Psychiatric Services, 54, McLeod, B. D., & Weisz, J. R. (2005). The therapy process 1544–1546. doi:10.1176=appi.ps.54.11.1544 observational coding system alliance scale: Measure characteristics Garland, A. F., Brookman-Frazee, L., Hurlburt, M. S., Accurso, E. and prediction of outcome in usual clinical practice. Journal of C., Zoffness, R. J., Haine-Schlagel, R., & Ganger, W. (2010). Mental Consulting and Clinical Psychology, 73, 323–333. doi:10.1037= health care for children with disruptive behavior problems: A view 0022-006X.73.2.323 inside therapists’ ofﬁces. Psychiatric Services, 61, 788–795. McLeod, B. D., & Weisz, J. R. (2010). The therapy process doi:10.1176=appi.ps.61.8.788 observational coding system for child psychotherapy strategies scale. Garland, A. F., Hough, R. L., McCabe, K. M., Yeh, M., Wood, P. A., Journal of Clinical Child & Adolescent Psychology, 39, 436–443. & Aarons, G. A. (2001). Prevalence of psychiatric disorders in doi:10.1080=15374411003691750 youths across ﬁve sectors of care. Journal of the American Academy Miklowitz, D. J., & Goldstein, M. J. (1997). Bipolar disorder: of Child & Adolescent Psychiatry, 40, 409–418. doi:10.1097= A family-focused treatment approach. New York, NY: Guilford. 00004583-200104000-00009 Miller, W. R., & Rollnick, S. (1991). Motivational interviewing: Prepar- Harpaz-Rotem, I., Leslie, D., & Rosencheck, R. A. (2004). Treatment ing people to change addictive behavior. New York, NY: Guilford. retention among children entering a new episode of mental health Morrissey-Kane, E., & Prinz, R. J. (1999). Engagement in child and care. Psychiatric Services, 55, 1022–1028. doi:10.1176=appi.ps. adolescent treatment: The role of parental cognitions and attri- 55.9.1022 butions. Clinical Child and Family Psychology Review, 2, 183–198. Hoagwood, K. E., Cavaleri, M. A., Olin, S. S., Burns, B. J., Slaton, E., doi:10.1023=A:1021807106455 Gruttadaro, D., & Hughes, R. (2010). Family support in National Research Council and Institute of Medicine. (2009, March children’s mental health: A review and synthesis. Clinical Child 12). Preventing mental, emotional, and behavioral disorders among and Family Psychology Review, 13, 1–45. doi:10.1007=s10567- young people: Progress and possibilities (O’Connell, M. E., Boat, 009-0060-5 T., Warner, K. E.; Committee on the Prevention of Mental Jorm, A. F. (2000). Mental health literacy: Public knowledge and Disorders and Substance Abuse Among Children, Youth, and beliefs about mental disorders. The British Journal of Psychiatry, Young Adults: Research Advances and Promising Interventions; 177, 396–401. doi:10.1192=bjp.177.5.396 Board on Children, Youth, and Families, Division of Behavioral Karver, M. S., Handelsman, J. B., Fields, S., & Bickman, L. (2006). and Social Sciences and Education). Washington, DC: The National Meta-analysis of therapeutic relationship variables in youth and Academies Press. family therapy: The evidence for different relationship variables Nelson, T. D., Steele, R. G., & Mize, J. A. (2006). Practitioner in the child and adolescent treatment outcome literature. Clinical attitudes toward evidence-based practice: Themes and Psychology Review, 26, 50–65. doi:10.1016=j.cpr.2005.09.001 challenges. Administration and Policy in Mental Health and Kataoka, S. H., Zhang, L., & Wells, K. B. (2002). Unmet need Mental Health Services Research, 33, 398–409. doi:10.1007= for mental health care among U.S. children: Variation by s10488-006-0044-4

ethnicity and insurance status. American Journal of Psychiatry, Nock, M. K., & Ferriter, C. (2005). Parent management of attendance 159, 1548–1555. doi:10.1176=appi.ajp.159.9.1548 and adherence in child and adolescent therapy: A conceptual and Kazdin, A. E. (1996). Dropping out of child psychotherapy: Issues for empirical review. Clinical Child and Family Psychology Review, 8, research and implications for practice. Clinical Child Psychology and 149–166. doi:10.1007=s10567-005-4753-0 Psychiatry, 1, 133–156. doi:10.1177=1359104596011012 Nock, M. K., Ferriter, C., & Holmberg, E. (2007). Parent beliefs about Lukens, E. P., & McFarlane, W. R. (2004). Psychoeducation treatment credibility and effectiveness: Assessment and relation as evidence-based practice: Considerations for practice, research, to subsequent treatment participation. Journal of Child and Family and policy. Brief Treatment and Crisis Intervention, 4, 205–225. Studies, 16, 27–38. doi:10.1007=s10826-006-9064-7 doi:10.1093=brief-treatment=mhh019 Nock, M. K., & Kazdin, A. E. (2001). Parent expectancies for child Malik, M. L., Beutler, L. E., Alimohamed, S., Gallagher-Thompson, therapy: Assessment and relation to participation in treatment. D., & Thompson, L. (2003). Are all cognitive therapies alike? Journal of Child & Family Studies, 10, 155–180. A comparison of cognitive and noncognitive therapy process and Nock, M. K., & Kazdin, A. E. (2005). Randomized controlled trial implications for the application of empirically supported treatments. of a brief intervention for increasing participation in parent Journal of Consulting and Clinical Psychology, 71, 150–158. management training. Journal of Consulting and Clinical Psychology, doi:10.1037=0022-006X.71.1.150 73, 872–879. doi:10.1037=0022-006X.73.5.872 Martinez, J. I., Lau, A. F., & Chorpita, B. F. (2012). Psychoeducation Owens, P. L., Hoagwood, K., Horwitz, S. M., Leaf, P. J., Poduska, and engagement strategies observational coding system. Unpublished J. M., Kellam, S. G., & Ialongo, N. S. (2002). Barriers to children’s coding manual. mental health services. Journal of the Academy of American Child McCabe, K. M. (2002). Factors that predict premature termination & Adolescent Psychiatry, 41, 731–738. doi:10.1097=00004583- among Mexican-American children in outpatient psychotherapy. 200206000-00013 Journal of Child and Family Studies, 11, 347–359. doi:10.1023= Pavuluri, M. N., Graczyk, P. A., Henry, D. B., Carbray, J. A., A:1016876224388 Heidenreich, J., & Miklowitz, D. J. (2004). Child- and family- McKay, M. M., Harrison, M. E., Gonzales, J., Kim, L., & Quintana, focused cognitive-behavioral therapy for pediatric bipolar disorder: E. (2002). Multiple-family groups for urban children with conduct Development and preliminary results. Journal of the American difﬁculties and their families. Psychiatric Services, 53, 1467–1468. Academy of Child & Adolescent Psychiatry, 43, 528–537. doi:10.1176=appi.ps.53.11.1467 doi:10.1097=00004583-200405000-00006 McKay, M. M., Nudelman, R., McCadam, K., & Gonzales, J. (1996). Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling Evaluating a social work engagement approach to involving inner-city strategies for assessing and comparing indirect effects in multiple children and their families in mental health care. Research on Social mediator models. Behavior Research Methods, 40, 879–891. Work Practice, 6, 462–472. doi:10.1177=104973159600600404 doi:10.3758=BRM.40.3.879 McKay, M. M., Stoewe, J., McCadam, K., & Gonzales, J. (1998). Prinz, R. J., & Miller, G. E. (1994). Family-based treatment for Increasing access to child mental health services for urban children childhood antisocial behavior: Experimental inﬂuences on dropout PSYCHOEDUCATION AND PARENT ENGAGEMENT 15

and engagement. Journal of Consulting and Clinical Psychology, 62, Mental Health. (2012). Testing standard and modular designs 645–650. doi:10.1037=0022-006X.62.3.645 for psychotherapy treating depression, anxiety, and conduct Shirk, S. R., & Saiz, C. C. (1992). Clinical, empirical, and problems in youth: A randomized effectiveness trial. Archives of developmental perspectives on the therapeutic relationship in child General Psychiatry, 69, 274–282. doi:10.1001=archgenpsychiatry. psychotherapy. Development and Psychopathology, 4, 713–728. 2011.147 doi:10.1017=S0954579400004946 Weisz, J. R., & Kazdin, A. E. (Eds.). (2010). Evidence-based Szapocznik, J., Perez-Vidal, A., Brickman, A. L., Foote, F. H., & psychotherapies for children and adolescents. New York, NY: Santisteban, D., Hervis, O., & Kurtines, W. M. (1988). Guilford. Engaging adolescent drug abusers and their families in treatment: Yeh,M.,McCabe,K.,Hough,R.L.,Lau,A.,Fakhry,F.,& A strategic structural systems approach. Journal of Consulting Garland, A. (2005). Why bother with beliefs? Examining and Clinical Psychology, 56, 552–557. doi:10.1037=0022-006X. relationships between race=ethnicity, parental beliefs about 56.4.552 causes of child problems, and mental health service use. Journal Weisz, J. R., Chorpita, B. F., Palinkas, L. A., Schoenwald, S. K., of Consulting and Clinical Psychology, 73, 800–807. doi:10.1037= Miranda, J., Bearman, S. K., ... the Research Network on Youth 0022-006X.73.5.800